CT perfusion stroke lesion threshold calibration between deconvolution algorithms
Bibliographic record
Abstract
Structured Abstract Objective CTP is an important diagnostic tool in managing patients with acute ischemic stroke, but challenges persist in the reliability of stroke lesion volumes determined with different software. We investigated a systematic method to calibrate CTP lesion thresholds between deconvolution algorithms using a digital perfusion phantom. Approach The accuracy of one model-independent and two model-based deconvolution algorithms in estimating ground truth cerebral blood flow (CBF) and Tmax in the phantom was quantified. Reference thresholds for ischemic core and penumbra were model-independent CBF<30% and Tmax>6 s, respectively, which is the current clinical standard. The equivalent model-based CBF and Tmax thresholds were determined by comparing linear regressions of phantom ground truth and deconvolution-estimated perfusion between algorithms. Calibrated thresholds were then validated in 63 patients with large vessel stroke by comparing admission CTP ischemic core and <3-hour diffusion-weighted imaging (DWI) lesion volume by Bland-Altman analysis. Agreement in target mismatch (core < 70 ml, penumbra ≥ 15 ml, mismatch ratio ≥ 1.8) determined by the three methods was assessed by Cohen’s kappa (κ) and concordance. Main Results The calibrated thresholds were CBF<15% and Tmax>6 s for both model-based methods. DWI minus CTP lesion mean volume differences (95% limits of agreement) were +16.2 (−30.9 to 63.3) ml, +10.9 (−32.9 to 54.7) ml, and +13.8 (−48.1 to 75.7) ml for model-independent and the two calibrated model-based approaches, respectively. Agreement in mismatch profiles with the two model-based deconvolution methods versus model-independent assessment was κ = 0.87 (95% confidence interval [CI]: 0.72 to 1.00) and κ = 0.86 (95% CI: 0.70 to 1.00), and both achieved 95% concordance. Significance We reported a systematic method of calibrating perfusion thresholds between deconvolution algorithms based on their quantitative accuracy. This may harmonize ischemic lesion volumes determined by different CTP software.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.009 | 0.054 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".